[Daniel, Torrey]

We've been running a test for a few weeks tracking the efficiency of the power distribution center. In the current configuration, it is possible to get the efficiency of the fiber collimator + fibers to high 80 percents. It should be noted as well that there were large temperature changes in the first three days of this data.

The MISC path is a standard 2 axis tip tilt mount compared to the other 3 paths which are the 5 degrees of freedom mounts. The initial theory was there is a relaxation time to some of these mounts on a several day time scale leading to loss of power in the 5 DOF mount. But this suggests something drifting upstream to me.

Attached are descriptions of the measurement setup used in log post 11449, as well as transfer function and noise data.

The piezo used here was Thorlabs, as described in 11449. The piezo transfer functions taken in log post 11373 were of the Noliac, would it be useful to take the same measurement of the Thorlabs piezo? Should it give us the same transfer function we will get by quotienting the open loop transfer function by the controller?

The noise measurements were taken with the laser locked, with the fan off and fan on.

Continuation of 11447.

Attached are several different OLG measurements with the cavity locked. The differences are as follows:

Thorlabspiezo-differentsm1tightness.png is the thorlabs piezo on both traces, without adjusting the controller shape. The active trace is screwing in the SM1 ring a fair amount, almost no slack left.

noliacpiezodifferentsm1tighness.png is the Noliac piezo on both traces. Additionally without changing the loop shape (between the two measurements, but changed from piezo to piezo). The active trace is again screwing in the SM1 ring a good amount.

Differentpiezos.png is a comparison of the two piezos with the same loop shape. Note the large difference in UGF.

Take aways:

1) Tightening the SM1 ring doesn't seem to effect the OLG when using the thorlabs piezo. Therefore I'm not sure its worth testing the stiffnesses of the o-rings used in the set up for the thorlabs piezo. For the noliac, it does seem to flatten some lower frequency parts of the OLG. It also lowered the UGF.

2) The 3.3 kHz mystery feature is in the transfer function of both piezos. If this feature is eliminated, I think the thorlabs piezo could become sufficient for locking these cavities with a mirror.

I think we should buy some more of the thorlabs piezos at different sizes, as they are relatively cheap, and test them on this cavity. 

[Daniel, Torrey]

We replaced the Noliac NAC2125-H08 piezo with a Thorlabs PA44M3KW to see if it changed the UGF. Everything else stayed the same physically, the control loop of the slow controller was shaped to optimize lock quality. The UGF went from 1 kHz to 3.5 kHz. Visibily the lock seems much more stable. I think it is worth testing difference stiffnesses of the viton o-rings in the piezo mirror set ups (the piezo pushes into the viton o-ring to change the length of the cavity). 

Take the custom made square aluminum base with four 1/4-20 holes, screw in an SM1 ring, then add a viton o-ring and mirror face down. With the custom made top, which looks like two concentric cylinders with a slot, add 3 #4 nylon tipped set screws. Add a tested piezo (see the previous piezo testing posts) with the wires through the slot. Center the piezo by gently tightening the set screws. Take the piezo base and back off the mirror so that it is under the top face. Add the top with the piezo to the base, and insert and fully tighten the 1/4-20 screws. Hold the assembly upright gently screw in the sm1 ring until you feel some resistance due to the piezo interfacing with the mirror. Then back off the set screws that hold the piezo.

<I will add photos the next time I do this procedure>

I reran the latest simulation with the same parameters as (https://mccullerlab.com/logs/lab/index.php?callRep=11430) but with a fused silica coating material instead of silicon. The Q of the coating remained 1e4. The noise floor from the coating goes up, but the qualitative behavior is very similar as silicon coatings with the same loss.

Potential reason for the loss in power over time for the fibers on the power distribution center: This says:

Wide-Key-Slot Mating Sleeves - While this configuration is acceptable for patch cables with FC/PC connectors, for FC/APC applications, we recommend using narrow-key-slot mating sleeves to ensure optimum alignment. Our patch fibers are FC/APC. Additionally the adjustable fiber collimators we purchased (pdf attached) are wide key. They don't sell the collimator as narrow key, but we could potentially request it.

[Torrey, Daniel, Sander]

While trying to lock a readout filter cavity using a setup that enabled reliable and stable locking before, we noticed mode hops (i.e. we saw different modes flashing in the cavity in random succession) during scanning and locking. There was also a much larger drift of the error signal than previously observed. I suspected this could be due to mode hops in the seed laser, caused by back-reflection into the seed laser. A seed laser output beam pick-off after the fibre-PBS is currently used as input for a Michelson setup, and this output has no Faraday isolator to prevent back reflections. Blocking this output with a beam dump resolved all problems and enabled locking of the filter cavity as before. 

Going forward we should install a Faraday isolator at the fibre output or only pick off beams after the amplifier's Faraday isolator.

I set up a Michelson Interferometer with a 17 inch Schnupp Asymmetry to test the noie from the flexure mount. A 750 mm focal length lens is added 6 inches away from the long arm end mirror so that the beams are approximately the same size at the beam splitter. The visability is approximately 50% without alignment to maximize this.

Before the input port, I added a 10:90 (R:T) BS and photodiode to act as a power reference. When the laser's frequency is modulated, the power is also modulated. By dividing the Michelson output power by this reference power, the direct frequency dependence can be seen. This will allow us to put the Michelson on a midfringe and to test the noise from the flexure mount.

[Torrey, Daniel]

We set up a Michelson interferometer with a 14 inch Schnupp Asymmetry to lock the output to midfringe. We added a lens so that the beams are the same size at the beamsplitter and roughly the same Guoy phase. However, modulating the laser's frequency also modulates its power. Next week, we will set up a photodiode with a reference power before the beamsplitter and divide the output with the input to see if the fringe changes.

Locked the cavity with the DC modulation port of the ULN15TK laser. Couple take aways:

-The UGF is pushed to ~2kHz (was ~1kHz).

-Some very strong peaks introduced via this port at ~830 Hz, 1.6 kHz, etc...

-the lock is more resilient against sound, but not immune.

I set up a piezo assembly with a 10D20DM.8 1550 nm HR mirror, a viton o-ring, and a Noliac NAC2125-H08 piezo. This assembly is mounted on a 5-axis stage so that it can be aligned . I moved the flexure mount assembly to a new, fixed mount since the flexure mount itself can be moved. Once the mirrors are aligned, the michelson should be able to be locked. This should hopefully provide more sensitivity for our noise test.

I am setting up a second Noliac NAC2125-H08 for the Michelson and also eventually for the 2nd output filter cavity.

Capacitance: 2.3684 uF

Resistance: 4.132 kOhm

When I press on the piezo, there is a positive voltage when I assign the lead with the black dots to be the positive lead.

I have flagged this lead with kapton tape.

[Daniel, Sander, Torrey]

We wanted to see if we could lock our michaelson in the lab using the frequency of the laser. However, the external modulation ports on the turnkey laser we have has an input voltage requirement of +/- 5V. The manual says this modulates the current at 2 mA/V. The graphs on thorlabs say changing the current by ~20mA should correspond to changing the wavelength of the laser light by order 10 pm. If our interferometer has interference fringes lambda/2 = 775 nm apart and we can only modulate by about 10 pm maximum, we will never be able to lock a michaelson using the laser frequency alone.

We were able to inject a 150 kHz signal into the AC modulation port of the laser and can directly read that out on a PD, but only by blocking one of the arms of the michaelson, suggesting this is purely modulating the amplitude and not frequency. We also tried using the temperature unsuccessfully, which does not have specs online of its mA/V response.

I using Levin's method (Internal thermal noise in the LIGO test masses: A direct approach) of using the suceptability of a mirror to a force resembling the laser beam profile to model the thermal noise, I modeled the GQuEST end mirrors as 2 mm thick, 1 in side length squares made of silicon with a Q of 10^6. I added a coating 20 um thick made of silicon as well but with a Q of 10^3. I constrained the 4 side of the mirror to only move in the z-axis. The spot radius is 2 mm. The face of the mirror had 52 x 52 mesh points, the substrate had 10 layers of mesh, and the coating had 3 layers of mesh. I simulated from 500 kHz to 16.5 MHz with 100 kHz frequency spacing. It look 5 hours and 17 minutes. Attached is the data converted into an ASD. The coating thermal noise, due to its low Q and that the force is applied on it, dominates the "bottom of the bucket".